Tada - all good points.

I think your first option might be possible in the case of CHF readmissions. The FDA might ask for additional evidence of improvement in heart function (eg, cardiac output, contractility, etc). A clinical study on that would be a very much smaller undertaking than a specific CHF readmissions trial.  

I agree a bolt-on would have to use the same selection criteria, endpoints, etc, and would probably be quite a big undertaking.

A clinical endpoint trial to confirm the apparent anti-diabetic combo effect would also be a big undertaking.

I cannot see too many options for a quick follow-on study unless there was evidence of improvement in eGFR. Reproducing and/or expanding on that might not be too big an undertaking.